A Tap Tempo Control Chip

[MoltenVoltage]

Here are the PT2399 resistances charted against the delay time:



As you can see, an equal change in resistance does not create an equal change in delay.  A good example of this:
- going from 7200 to 8200 ohms (1000 ohm increase) increases delay by 10.4 ms
- going from 8200 to 9200 ohms (1000 ohm increase) increases delay by 12.5 ms

what's more, the delay times do not evenly increase as resistance goes up:
- going from 2400 to 3400 ohms (1000 ohm increase) increases delay by 11.5 ms!

This information is straight off the Princeton PT2399 datasheet

[The Tone God]

Is there any reason why a tap tempo control chip cannot be adapted to controlling a parameter other than delay time or tremolo speed?

Of course but it becomes the same issue, how to translate the time value derived to some type of appropriate output.

And for everyone's enjoyment here is a graph of resistance vs. delay time just based on the data in the PT2399 datasheet.



In case it is not obvious that is NOT linear or even log.

You can do some things with this IC though like a chopper or maybe a metronome.

...an extensive table which is not possible with an 8 pin microcontroller.

I beg to differ.

Andrew

[MoltenVoltage]

A standardized table that would provide delay time accuracy equal to that of MV-52 (better than 1/1000th of a second) with the PT2399 is simply not possible with an 8 pin microcontroller.  You say that you "beg to differ".  If you know something we don't, please share!  Keep in mind that you would need to test the resistance v. delay time every 1ms, log that resistance then program the table to assign the proper value.  Over 400ms, that would be 400 entries in the table.  Even if you did all that, you would still need to tweak your table based on a particular chip's idiosyncrasies (not to mention any external components), making a standardized table impossible.

As far as applications for our Universal Tap Tempo Control Chip, they are more than just "metronomes and chopper effects".  The most obvious example, as shown by our Application Note 1, is a sequencer design, but many other designs are possible if you let yourself think outside the box.   

http://www.smallbearelec.com/Projects/MV-52App.pdf

[The Tone God]

A standardized table that would provide delay time accuracy equal to that of MV-52 (better than 1/1000th of a second) with the PT2399 is simply not possible with an 8 pin microcontroller.  You say that you "beg to differ".  If you know something we don't, please share!  Keep in mind that you would need to test the resistance v. delay time every 1ms, log that resistance then program the table to assign the proper value.  Over 400ms, that would be 400 entries in the table.  Even if you did all that, you would still need to tweak your table based on a particular chip's idiosyncrasies (not to mention any external components), making a standardized table impossible.

Perhaps I should have been alittle clearer as I thought the quote I made was descriptive enough. The main point I was disputing was that it was not be possible to have such a large table in a 8 pin microcontroller (uC). This maybe true for your chosen uC but it is still possible within the world of uCs to have such a large table in a 8 pin microcontroller. You do bring some other points up so I'll throw some other thoughts out.

Who said one needs to test for every 1ms ? The limitation in this application will not be the accuracy of uC derived time but the accuracy of the resistive element you choose to control the PT2399 with. Even if you had 1ms accurate tables there are very few resistive elements that would provide matching accuracy. Thus as I said earlier:

If you are willing tolerate alittle bit of inaccuracy then you can get something working within reasonable resources.

Princeton does not publish the clock tolerances and AFAIK no one has publicly published data on this so we are all guessing at this point at how accurate the clock is. The clock variation maybe within the level of tolerance.

Andrew

[StephenGiles]

Please forgive my cynicism but in a band situation, tap tempo may be quite useless because one is required to play guitar rather than play effects. But don't let me stop the infinitely entertaining thread

[Mark Hammer]

Mid-song, I'd agree with you, but many band scenarios allow the guitar player to think of the tempo they want, tap a bit before a song starts, or maybe even tap into the lead-in of a song, to establish the tempo, and the band joins in.  Not always, mind you, but it happens.  No different than a drummer tapping their sticks together to set the tempo for everyone else.

I still find tap-tempo rather pointless for long sweeps on flangers or phasers, but that's me.  YMMV.

[ianmgull]

Please forgive my cynicism but in a band situation, tap tempo may be quite useless because one is required to play guitar rather than play effects. But don't let me stop the infinitely entertaining thread

I've heard people say this in other threads and I still can't wrap my head around it. Isn't the idea of a tap tempo that you use your foot to do the tapping?

I have quite a few songs where I make use of synchronized delays. This being the case I have to readjust my tempo a couple times in a song to keep in line with peoples natural variations.

I have a couple Line 6 Echo Parks that I have linked up to a common tempo switch. I run them in parallel each with a different setting. If you're into the kind of rhythmic delay that U2 is known for tap temp delays are pretty much a necessity live.  It's really not that hard. If you can tap your foot while you play you can use a tap tempo delay in this way.

[The Tone God]

Yes tap tempo is only really useful for quicker stuff which is why you don't see tap tempo chorus pedals. I still think we are jumping around the accuracy issue. Do players really need a 1ms accurate delay ? Most don't and I would bet in most band situations the audience will not notice the delay being slightly off. Those who do need this level of accuracy will shift to a commercial digital delay so that issue moves out of the scope of this discussion.

We should also be alittle bit honest about the PT239x. It is a cheap and easy to use IC which is great for cheap commercial delays and us DIYers. It was never meant to be a highly accurate delay so to argue about 1/1000th of a second accuracy is kind of a waste. You'll never get it. I think most DIYers would be happy just to get the delay into the rough average time wise.

As far as applications for our Universal Tap Tempo Control Chip, they are more than just "metronomes and chopper effects".  The most obvious example, as shown by our Application Note 1, is a sequencer design, but many other designs are possible if you let yourself think outside the box.   

I forgot to respond to this previously. Firstly just make sure my tone is understood please take this as a friendly suggestion to help with your product. No long lasting company produces a product and then expects people to find uses for it. If you are in the business of selling a product it is your job to tell the potential buyers, like members of this forum, what we can and cannot do with your product. You should not tell potential buyers to "think outside the box" and then offer no suggestions other then one application which I think we already acknowledged earlier in this thread. If you want to sell more product I would suggest your company come up some other ideas for the use of this product along with detailed app. notes for those uses. Making it as easy as possible for buyers to use your product will make more people want to buy your products.

Andrew

[R.G.]

Just  curious.

How many people is "we" in Molten Voltage?

[Processaurus]

Really neat chip to have available to the average stompboxing type that doesn't know how to program.  I'll definitely be getting a few, have a music from outer space CV sequencer PCB lying around and that would be outstanding.

Questions for Molten Voltage, does tapping the switch once get the clock back on beat when it starts to drift?  Does the tap tempo switch need to be debounced?  How many taps does it take to set the tempo, and when does the tempo change to the one tapped in?  What happens if you keep tapping the pedal, does it average the tap time, or just go with the last taps?

[Mark Hammer]

Re: Usefulness.

When a drummer taps their sticks together, or when someone counts off a Ramones-like "1-2-3-4!!", they are setting the approximate tempo of the song.  There is no assumption about precision, but more about average feel.  I'm pretty certain that the actual BPM will drift over time but the musicians will generally "regress toward the mean", in the sense that their average pace will coalesce as the tune moves on (or fall apart if it was too damn fast for them in the first place).  Whether it started off a little too fast or a little too slow is not always that important.  The main thing is it's in the ballpark, tempo-wise.

And that's pretty much what one expects from tap tempo: the right feel.

Which brings me to another issue:  Is the feel most appropriately achieved or "sensed" via 2 taps, or do you need more to nail it?  By that, I don't mean whether it is technically possible to provide a sensed tempo from two taps, because this chip clearly demonstrates that it is.  What I want to know is whether the sensed tempo would be more "correct" with 3 or 4 than with 2 taps.  Do people who use 2 taps find that what they got was exactly what they were aiming for?

And another thing.  If you didn't get the tempo you were aiming for, how do you establish a new tempo?  Do you simply tap 2 more times and if so, after what minimum interval?  Just trying to get a handle on how to plan for this chip, because dammit...it's interesting.

[trendyironicname]

..Do players really need a 1ms accurate delay ? Most don't and I would bet in most band situations the audience will not notice the delay being slightly off...

My feet are retarded.  Would need my drummer to do it for me.  I'd prefer something to quantize my taps . "I know you're doing this but here's what you really want.."

[The Tone God]

Which brings me to another issue:  Is the feel most appropriately achieved or "sensed" via 2 taps, or do you need more to nail it?  By that, I don't mean whether it is technically possible to provide a sensed tempo from two taps, because this chip clearly demonstrates that it is.  What I want to know is whether the sensed tempo would be more "correct" with 3 or 4 than with 2 taps.  Do people who use 2 taps find that what they got was exactly what they were aiming for?

I have used delays with the ability to average more then two taps and I find after the initial two taps you don't get much more out of the extra taps. The only time I found it useful is when my foot slipped off the switch I was able to recover the tempo but to redo two taps is not that big of deal.

My feet are retarded.  Would need my drummer to do it for me.  I'd prefer something to quantize my taps . "I know you're doing this but here's what you really want.."

Computers can only do so much. GIGO ?

Andrew

[MoltenVoltage]

Where to begin...

First, while we agree that 1ms accuracy isn't that big of a deal playing live if you have a "flexible" drummer, accuracy is a big issue if you are playing along with other tracks while overdubbing.

Assume a typical song of 120 bpm (2 beats per second : 500 ms between beats)
If this is a 3 minute song, it is 180 seconds (180,000 ms) long.

Now assume you want the effect tempo at 1/4 notes, you are at 125ms between quarter notes.

If your effect gets out of time 1ms per second, by the end of the song (180 seconds), you are well over a quarter note out of time (180ms).

Instead of being on the downbeat, you are well past the snare beat!

With MV-52, the tempo is set when you tap twice.  You can tap in a new tempo at any time after about half a second.

As far as the suggestion that we should provide piles of design ideas, we are also in the pedal manufacturing business and spend a lot of resources copyrighting and patenting our intellectual property.  The application note we provided is very detailed and, with just a little creativity, allows builders to create all kinds of interesting sequenced effects.  We are sorry if this is not good enough for you, but we aren't going to give away our future pedal designs for the sake of selling this chip.  We are confident that there are builders out there who can create original designs and will see the potential in our Universal Tap Tempo Control.  (see Processaurus' comment, above)

Other questions:

1) Does tapping the switch once get the clock back on beat when it starts to drift?

- No, but tapping twice does

2) Does the tap tempo switch need to be debounced?

- No, it is internally debounced.  Check out the datasheet if you want to see how to get started with just a few basic components:
http://www.smallbearelec.com/Projects/MV-52Data.pdf

3) How many taps does it take to set the tempo, and when does the tempo change to the one tapped in?

- 2 taps always sets a new tempo, and as soon as you tap the second time, the new tempo begins

4) What happens if you keep tapping the pedal, does it average the tap time, or just go with the last taps?

- If you keep tapping, the tempo will be the time between the last two taps (assuming an even number of taps)

Hopefully we caught all the questions.  Thanks for all the feedback!

[trendyironicname]

Computers can only do so much. GIGO ?

Indeed, Gigo.  Indeed. 

Must've heard my stuff..


 

[R.G.]

Does the tap tempo switch need to be debounced?

*All* switches used by digital logic for inputs need to be debounced. Sometimes people do a less good job of this, with mildly amusing to hilarious results. 

[trendyironicname]

   I kind of want one of the chips.

[Taylor]

Where to begin...

First, while we agree that 1ms accuracy isn't that big of a deal playing live if you have a "flexible" drummer, accuracy is a big issue if you are playing along with other tracks while overdubbing.

Assume a typical song of 120 bpm (2 beats per second : 500 ms between beats)
If this is a 3 minute song, it is 180 seconds (180,000 ms) long.

Now assume you want the effect tempo at 1/4 notes, you are at 125ms between quarter notes.

If your effect gets out of time 1ms per second, by the end of the song (180 seconds), you are well over a quarter note out of time (180ms).

Instead of being on the downbeat, you are well past the snare beat!

This doesn't make sense to me. This would only be a problem if you let the same note repeat with 100% feedback for the entire song. As long as you are putting new notes into the delay line, each new note "resets" the beginning of the bar.

[MoltenVoltage]

This doesn't make sense to me. This would only be a problem if you let the same note repeat with 100% feedback for the entire song. As long as you are putting new notes into the delay line, each new note "resets" the beginning of the bar.

Good point.  With a basic delay, you really don't need as much accuracy.

When I wrote the last post, I was thinking about sequencer and chopper designs, where you need a super-accurate tap tempo control or you get out of time very quickly.  Maybe the PT2399 thing is do-able with a microcontroller since you just need to be in the ballpark with a delay...

[MoltenVoltage]

Does the tap tempo switch need to be debounced?

*All* switches used by digital logic for inputs need to be debounced. Sometimes people do a less good job of this, with mildly amusing to hilarious results. 

I think what Processaurus was asking was whether or not you need external circuitry to debounce the chip.  In this case, you don't because we did it with code and tested it using this standard momentary on pushbutton switch:
http://www.smallbearelec.com/Detail.bok?no=672

As you can see from the datasheet, the lower limit for the real-time tap interval is about .175 seconds (342 bpm):
http://www.smallbearelec.com/Projects/MV-52Data.pdf

This was necessary to prevent false switching.  By grounding pin 4 you can scale the tap interval to a minimum interval of .044 seconds.  This equates to 1,363 bpm.

MV-52's maximum possible interval between taps is 4.19 seconds (real-time).